topic RE: Reading underlined text with Decipher IDP- Challenges in Product Forum

Reading underlined text with Decipher IDP- Challenges

KrishnaElapavul — Wed, 02 Feb 2022 13:12:00 GMT

Hi
I am trying to read an underlined text through Decipher IDP. I am not getting the text what is expected , Any one can throw light how to tackle this scenario
here expected value is 4128.126.020, but it is adding dots between each letter as shown in the attachment /inserted image

------------------------------
Krishna Elapavuluri RPA Solution Lead
TEchnology Consultant
DXC.technology
Asia/Kolkata
------------------------------

RE: Reading underlined text with Decipher IDP- Challenges

Ben.Lyons1 — Wed, 02 Feb 2022 14:23:00 GMT

Hi Krishna,

Have you tried using any format expressions? This might help if the values are quite consistent in structure. e.g. [0-9]{4}\.?[0-9]{3}\.?[0-9]{3}

Thanks

------------------------------
Ben Lyons
Product Consultant - Decipher Specialist
Blue Prism
UK
------------------------------

RE: Reading underlined text with Decipher IDP- Challenges

KrishnaElapavul — Wed, 02 Feb 2022 14:32:00 GMT

Thanks Ben for your input

But structure is not consistent , different combinations will come at that position

Alphanumeric with different signs like . - /...

Thanks

Krishna

SKF

DXC Technology

T +91 99000 63042
Krishna-mohan.elapavuluri@dxc.com

DXC.com | Twitter* | Facebook* | LinkedIn* | Campaign Promotion*

Planned Leave: 23^rd Feb

Company Holidays: 1^st Mar.

RE: Reading underlined text with Decipher IDP- Challenges

Ben.Lyons1 — Wed, 02 Feb 2022 14:35:00 GMT

That's a shame.

I think you're going to have a a tough time as the OCR engine will find it challenging to separate the fonts from the line. Have you tried excluding the character _ in the exclude values setting?

Thanks

------------------------------
Ben Lyons
Product Consultant - Decipher Specialist
Blue Prism
UK
------------------------------

RE: Reading underlined text with Decipher IDP- Challenges

SergeiEvdokimov — Mon, 07 Feb 2022 12:39:00 GMT

Good afternoon Ben.

Please tell me where to look for documentation for DFD (wildcards, functions and formulas).

I couldn't find this information in the documentation for Decipher.

------------------------------
Sergei Evdokimov
------------------------------

RE: Reading underlined text with Decipher IDP- Challenges

Ben.Lyons1 — Mon, 07 Feb 2022 13:49:00 GMT

Hi Sergei,

The Format Expression field uses Regex (Regular Expression), which is a common tool for defining string formats. As such it's not something we go into detail in our help.

However, you can find a lot of help online, here's a couple of website that you may find useful.

Regex101 - Expression builder and tester
RegexOne - Training & How To

Thanks

------------------------------
Ben Lyons
Product Consultant - Decipher Specialist
Blue Prism
UK
------------------------------

RE: Reading underlined text with Decipher IDP- Challenges

SergeiEvdokimov — Mon, 07 Feb 2022 16:35:00 GMT

Thanks, but it's not very clear.
For example, a screenshot from the Decipher tutorial:

I couldn't find SPLITGET functions on regex.

There is also a question on how exactly regex is used (perhaps which engine), as an example of an expression that does not match regex in Decipher:

------------------------------
Sergei Evdokimov
------------------------------

RE: Reading underlined text with Decipher IDP- Challenges

SergeiEvdokimov — Tue, 08 Feb 2022 08:51:00 GMT

Found it 🙂 You have completed the guide section:
bpdocs

------------------------------
Sergei Evdokimov
------------------------------

RE: Reading underlined text with Decipher IDP- Challenges

Ben.Lyons1 — Tue, 08 Feb 2022 11:45:00 GMT

Hi Sergei,

That's good.

Hopefully for clarification, the Format Expression field is only for validation purposes not for extraction. While Decipher will use it to help locate the field, it will be used in combination with the other parameters like 'Sample Header'. So if it finds the correct sample header and the value doesn't confirm to the expression, it will either exclude it or highlight it as invalid. (this will depend on previous training and confidence.)

You can use a formula to extract different parts of a string using the guide mentioned in your link and new for 2.1 are Regex based formulas. E.g. REGEXSPLITGET.

Thanks

------------------------------
Ben Lyons
Product Consultant - Decipher Specialist
Blue Prism
UK
------------------------------